1. Add counter to your data set so that you have a row number for each ID - I'm assuming you know how to add an enumeration variable,if not see the link.
2. For each ID extract what the record number is on 21July2020
3. From #2 create two new variables, start/end with the record number that would be 11 days prior and 311 days prior.
4. Join data from original data to start/end and calculate mean
This shows you Steps 2 to 4. Untested.
data record_ranges;
set have (where = (date='21Jul2020'd));
record_start = counter - 11;
record_end = counter - 311;
run;
proc sql;
create table want as
select ID, mean(variable) as desiredMean
from ind as t1
left join record_ranges as t2
on t1.id = t2.id
and t1.counter between t2.record_start and t2.record_end
group by ID;
quit;
The name is Reeza, not REZZA.
Click image to register for webinar
